Parrot 2.9.1 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | merge gsoc_nci |
Set by moderator on 2 November 2010.
00:01 dalek left 00:02 dalek joined 00:11 patspam left 00:23 patspam joined 01:17 cotto left, cottoo joined 01:41 kid51 joined
Infinoid sorear: hi 01:52
cottoo dukeleto, ping 01:59
02:10 dngor left, patspam left 02:13 whiteknight joined, masonkramer__ joined 02:25 patspam joined
kid51 msg whiteknight Can you set up your new blog so that it clones to planet.parrotcode.org? Thanks. 02:33
aloha OK. I'll deliver the message.
whiteknight kid51: I don't actually know how to do that right now 02:34
I guess Coke would know
02:34 dngor joined
kid51 cotto probably knows. he recently got his new blog cloning there. 02:35
whiteknight Coke put it there
I'll see if I can get in touch with ask or robert on #perl 02:37
I know I used to have commit access to the repository that ran planet.parrotcode.org,but I have long since forgotten all the credentials 02:38
cottoo msg dukeleto What time will work for you to do the git migration this Monday? 02:42
aloha OK. I'll deliver the message.
02:58 kid51 left
whiteknight is that the general timeframe? Monday? 03:08
03:09 davidfetter joined
sorear Infinoid: still here? 03:10
03:17 whiteknight left
Infinoid sorear: for the moment :) 03:24
03:28 lidden left
sorear Infinoid: I've split the running instance of dalek into two programs 03:33
the big scary botnix part is just running a tail -f now 03:34
the scrapers are writing into a queue file
so. 03:36
1. I can easily do push stuff now - just write into /home/drain/dalek-poller/queue
(chattr +a to prevent accidents)
2. the interesting part of dalek no longer requires a bot account to be set up or anything like that. I hope to see a bit more hacking on github.com/sorear/dalek-poller now :) 03:38
3. dalek-poller can be restarted without affecting dalek
04:29 patspam left
cottoo sorear, are you the one who's to blame for dalek posting about wiki changes now? 04:50
sorear cottoo: I'm the only person actually doing anything with dalek now, so, I guess? 04:51
I'll look into wiki changes tomorrow 04:52
cottoo sorear++ then 04:58
04:58 cottoo is now known as cotto
dalek rrot: r49793 | petdance++ | trunk (4 files):
adding compiler annotations to functions and args
 05:49
06:11 theory left 06:51 theory joined, theory left 07:27 bluescreen left 07:57 fperrad joined 10:26 whiteknight joined 10:28 Kulag left, masonkramer__ left
whiteknight good morning, #parrot 10:29
10:30 contingencyplan left 10:31 contingencyplan joined, contingencyplan left 10:33 masonkramer_ joined 10:45 Kulag joined 10:51 contingencyplan joined 10:55 Kulag left, Kulag joined 11:00 contingencyplan left 11:02 Kulag left, cognominal left 11:03 cognominal joined 11:07 Kulag joined 11:19 jan_ joined 11:20 Kulag left 11:22 jan left, jan_ is now known as jan 11:23 Kulag joined 11:29 Kulag left 11:32 jan_ joined 11:35 jan left, jan_ is now known as jan 11:50 spinclad left 12:38 M_o_C joined 13:13 PerlPilot left
whiteknight I cannot figure out where the -G command-line argument is handled 13:13
or the --no-gc variant
13:13 kid51 joined
whiteknight blah, nevermind. I find it as soon as I complain 13:15
kid51 ./src/longopt.c:88: { 'G', 'G', (OPTION_flags)0, { "--no-gc" } }, 13:21
./src/main.c:310: " -G --no-gc\\n"
13:23 PerlJam joined
Infinoid cotto, sorear: My guess is that dalek isn't polling wiki changes because the new machine it's currently running on is missing some CPAN modules. I talked with diakopter about getting access to that machine at one point, but it hasn't happened yet 13:25
whiteknight kid51: yeah, that's what I found. However, looking around, I can't find any place that actually implements the logic for it
kid51 Maybe it was never implemented. 13:26
whiteknight ah, no. I think I just found it. The whole thing is implicit 13:27
we always turn off GC in IMCC. But if the -G option is specified we never re-enable it 13:28
damnit. All this commandline processing is terribly implemented, and spread out all over the code base 13:31
I thought the goal of something like longopt was to prevent that
kid51 whiteknight: If some newcomer to Parrot were to read your blog posts and then ask, "Where can I find out the *current* Parrot API?", how would you respond? 13:35
whiteknight there really is no "current Parrot API"
All exported functions are treated as fair-game, and most functions are exported 13:36
kid51 Does that mean that any function in the source code labelled PARROT_EXPORT is part of the API? 13:38
whiteknight see, that's the problem: Nobody has ever made a decision about that 13:39
and just because the functions are available for use doesn't mean they were ever sanctioned as being some kind of "API"
it's a case where nobody ever made a decision, so users did for themselves 13:40
kid51 Is this document of any use: ./docs/dev/parrot_api.pod ?
Would we consider something like NQP to be part of the API?
whiteknight sort of, but not really 13:41
kid51 or PCT?
whiteknight I don't know. I sincerely don't. 13:43
The "API" is whatever interface we should be exposing and supporting under the deprecation policy
but then there are sections of it, like there's an API to libparrot (which I am focusing on), there's an API to the parrot executable (the CLI options), an API for libraries, etc
it all depends what is supported by the deprecation policy 13:44
but again, nobody ever made an active decision about it: We had a bunch of functions, then we accepted a deprecation policy, and now everything is up in the air 13:45
kid51 So, does that imply that a re-look at the deprecation policy is needed as part of the product management team's work? 13:48
whiteknight I hope so. I hate that policy up and down 13:51
kid51 And, in brief, the new policy should be ...? 13:52
whiteknight All API members are opt-in. Nothing is considered supported unless we've evaluated it's fitness and specifically approved it 13:53
that's my opinion of it, I would like to hear other ideas 13:54
but I know several people, Rakudo devs especially, have been unhappy with the policy in the past
dalek rrot: r49794 | jkeenan++ | trunk/docs/dev/parrot_api.pod:
Correct POD formatting errors.
 13:55
kid51 Rakudo devs have been unhappy with "current policy" .. or with what you just advocated as a new policy?
whiteknight with current policy 14:01
I don't want to put words in anybody's mouth. We should ask them before assuming that they are actually unhappy with it 14:02
14:07 M_o_C left 14:33 patspam joined 14:43 spinclad joined 14:50 jan left 14:57 preflex left 14:59 preflex joined 15:06 preflex left 15:10 preflex joined 16:07 lidden joined
dalek TT #396 closed by jkeenan++: 32/64-bit cpu in smolder report 16:37
TT #396: trac.parrot.org/parrot/ticket/396
16:52 lucian joined 17:08 M_o_C joined 17:10 kid51 left 17:15 contingencyplan joined 17:32 lucian left 17:43 NotFound_gf left 17:50 theory joined
tadzik when is the new NCI planned to land? 17:55
dalek TT #1842 created by plobsing++: Linking against libparrot not as documented 17:56
TT #1842: trac.parrot.org/parrot/ticket/1842
plobsing tadzik: it is more or less in a mergeable state ATM, but svn has eaten more babies than usual on that branch, so I'm waiting for git.
tadzik plobsing: cool. So it could be even tuesday or so? 17:57
plobsing it could be today, if you're willing to resolve 92 meaningless svn conflicts (which I am not) 17:58
tadzik sure thing
(that neither do I)
17:59 lucian joined
moritz speaking of git... I noticed that audrey has two nicks in the svn log 18:07
plobsing whiteknight: please consider TT #1842 as my application to be a part of the embedding api task force
moritz audreyt and autrijus -- joining them wouldn't be bad
plobsing tadzik: (re: expectations of a better NCI API expressed on #perl6) the new NCI is the exact same as the old NCI, just with the ability to build call frames dynamically. It specifically does not change the way structs are handled. 18:11
tadzik oh 18:12
so what does it change actually?
plobsing you can call arbitrarily signature-d functions. Before you were limited to a finite set provided by core parrot (you also had the ability to add others with parrot_nci_thunk_gen, but nobody did) 18:14
tadzik: what about ManagedStruct/UnManagedStruct are insufficient in this case? 18:15
tadzik plobsing: looks so. Yet, as jnthn says on #perl6, it's hard but possible to achieve 18:16
cotto moritz, thanks for noticing that. dukeleto would be the one to ping. I think your change in CREDITS hasn't been moved to git yet so it won't be too much extra work to update audreyt's entries. 18:34
whiteknight plobsing: application accepted, sight unseen 18:58
I hadn't realized linking against libparrot was a problem 19:01
plobsing I've got an embbed parrot in vim project perpetually on the back burner. that's one of the problems I've run into.
dalek tracwiki: v4 | whiteknight++ | ProductManagementTeam 19:03
tracwiki: +plobsing
tracwiki: trac.parrot.org/parrot/wiki/Product...ction=diff 19:04
plobsing the others I can think of off the top of my head are already well known: managing Parrot_Interp, and calling exit()
cotto dukeleto, ping 19:07
plobsing afk # shopping 19:11
19:13 patspam left 19:25 allison joined
allison virtual developer summit sometime today? 19:25
I didn't see any reminders go out, or any settled time
cotto neither 19:30
dukeleto suggested that doodle.com would be a good tool, but nobody took up the task to set up an event and make sure people scheduled their availability. 19:33
I'll do that now.
19:35 Kulag joined
cotto How nice. They don't care what characters I have in my password. That's becoming depressingly rare. 19:36
whiteknight We definitely *should* have a PDS sometime soon, but I think this is not the weekend
something informal is always cool, but we obviously didn't send out enough firm, advanced notice for anything "official" 19:37
cotto whiteknight, no. That wouldn't be a reasonable amount of notice.
allison yeah, not enough advance warning 19:40
maybe next sunday?
19:42 jsut_ joined
cotto I'm setting up the event now. 19:44
whiteknight cotto++
cotto doodle does a good job of allowing people to list the the times that they'll be available so we can pretty easily pick one that works for the largest number of people.
19:47 jsut left
cotto any idea of how long we should expect the meeting to be? 19:47
whiteknight at least two hours 19:48
cotto I'll say 3. I'm sure it could go on forever, but we should be able to hit the most important points within that time. 19:51
wfy?
whiteknight fine for me 19:52
cotto done 19:54
whiteknight link? 20:02
cotto sent to parrot-dev 20:03
doodle.com/q4nwdt9i5k929n2s 20:04
Apparently changing your availability is nyi. 20:12
whiteknight cotto: did you remove yourself? 20:13
because as soon as I saved my times, you disappeared
cotto Yes. I need to figure out my schedule first. 20:18
I didn't realize that you can't go back and change, though there may be some way I missed.
ok. It does let you edit entries. I just didn't see it before because it's a bit circuitous. Go to "edit an entry" on the poll page and it'll let you do what you like. 20:25
whiteknight I haven't gotten an email about it to the mailinglist 20:29
cotto I sent it but haven't seen anything either. 20:39
figuring out my schedule for the next couple months is turning out to be quite the yak shaving exercise 20:49
whiteknight why? busy? 20:56
cotto somewhat, but I also need to do a bunch of coordinating for some church stuff 20:57
whiteknight ah, gotcha.
cotto and the typical method of organization is "oh, I guess we should do that." unless someone is on the ball 20:58
20:58 patspam joined, patspam left
sorear Infinoid: it appears to be working? c.f. "tracwiki" announce a couple hours ago 21:22
Infinoid: this may be related to me having installed Crypt::SSLeay on the box yesterday though. 21:23
Infinoid sorear: Oh, great. :) 21:48
cotto: Anything else missing at this point?
cotto Infinoid, wrt dalek I'm happy as long as it follows tickets, the wiki and scrapes the appropriate wiki pages. 21:49
Infinoid Okay, thanks. Trac tickets and trac wiki were the two things I had been told were down, but they both have reported things today 21:50
cotto Great. Thanks.
Infinoid sorear++ 21:51
cotto If anyone here is interested in attending the Parrot Developer Summit, please mark your availability on Doodle: doodle.com/q4nwdt9i5k929n2s 22:02
I sent a message to parrot-dev but it doesn't seem to have hit the list yet.
plobsing whoa, that's a lot of possible times 22:05
22:15 bluescreen joined 22:28 masonkramer___ joined, masonkramer_ left
sorear cotto: could you explain this scraping thing? 22:39
I was actually planning to rip out the scraping, since it's completely broken and the dalek-conf.json approach seems to work much better
Infinoid: Did you catch what I said yestday about the changes I've made? 22:40
22:46 M_o_C left 23:05 arnsholt left 23:17 fperrad left 23:24 patspam joined, patspam left 23:29 janus joined 23:36 dngor left 23:42 kid51 joined
sorear pings cotto 23:52
whiteknight dec 4th or 5th are front-runners right now 23:54
23:56 PacoLinux left 23:59 dngor joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!